tridge at samba.anu.edu.au
Tue Mar 10 13:20:11 GMT 1998
> it entirely depends on what's suitable. swat, as andrew has mentioned, is
> 500 lines (cool!). if it's that elegant, it would be nice to keep in that
well, it's small because it hooks into the rest of the Samba code. I
wouldn't really call it elegant :-)
> it may be sufficient to write perl scripts around smbclient. or it may be
> suitable to call smbclient from swat. or to call the same functions that
> smbclient does directly from swat.
the latter is the way to go. I'd much prefer that we create functions
that do common things than start launching smbclient or any other prog
from within SWAT. Otherwise it will all become too fragile - one small
layout change to smbclient and things will break.
See for example the status code I just added to SWAT. It uses a utility
function in the locking code to allow easy enumeration of the current
locks. The amount of replicated code is minimal.
> but _first_, identify the needs and goals. then go for the design,
> working through to implementation. do a proper software engineering job,
> in other words.
crumbs! That would take all the fun out :-)
More information about the samba-technical